智源研究院测试11款大模型：AI智能体或正在改变生物安全风险边界
2026-07-22
每日经济新闻
The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.
A joint study by Beijing Academy of Artificial Intelligence and Peking University found that large language model intelligent agents can generate DNA/RNA sequences and experimental protocols that bypass biosecurity screening. Physical lab validation confirmed these AI-generated plans are executable, lowering the knowledge barrier for potentially dangerous biological synthesis and posing significant biosecurity risks.[AI generated]
Why's our monitor labelling this an incident or hazard?
The event involves the use of AI systems (large language model intelligent agents) whose outputs can directly undermine existing biosecurity safeguards, potentially leading to harm to communities and the environment through misuse of biological experimentation knowledge. Although no specific harm has yet occurred, the AI's capabilities plausibly lead to an AI Incident by changing the biosecurity risk boundary and enabling unsafe practices. Therefore, this constitutes an AI Hazard, as the article focuses on the plausible future harm and systemic vulnerabilities introduced by AI, rather than a realized incident.[AI generated]